4′-Thionucleosides are attractive compounds with respect to antiviral and antineoplastic activity. For example, 1-(2-deoxy-2-fluoro-4-thio-β-D-arabinofuranosyl)cytosine (4′-thio-FAC) was shown to have excellent antitumour activity in vitro and in vivo [Y. Yoshimura et al; J. Org. Chem. 1997, 62, 3140-3152; S. Miura et al, Cancer Lett. 1998, 129, 103-110; S. Miura et al, Cancer Lett. 1999, 144, 177-182; Y. Yoshimura et al, Bioorg. Med. Chem. 2000, 8, 1545-1558; D. A. Zajchowski et al, Int. J. Cancer 2005, 114, 1002-1009].
The invention relates in particular to a novel process for preparing compounds of formula I
in which R1 represents —C(O)—C1-C6-alkyl or —C(O)-aryl; and R2 represents C1-C6-alkyl, C1-C4-perfluoroalkyl or aryl.
The compounds of the general formula I are key intermediates in the preparation of 4′-thionucleosides.
1-(2-Deoxy-2-fluoro-4-thio-β-D-arabinofuranosyl)cytosine [4′-thio-FAC]:

There is a particular interest in the preparation of 1-(2-deoxy-2-fluoro-4-thio-β-D-arabinofuranosyl)cytosine (4′-thio-FAC) with a view to the compound II:
where in the text below the α-diastereomer of the anomeric acetate is referred to as IIa and IIβ is used for the n-isomer.
This compound and its preparation was described for the first time in WO 97/73993, WO 97/038001 and in the literature associated therewith [Y. Yoshimura et al, J. Org. Chem. 1999, 64, 7912-7920; Y. Yoshimura et al, Nucleosides Nucleotides 1999, 18, 815-820; Y. Yoshimura et al, Nucleic Acids Symposium Series 1998, 39, 11-12; Y. Yoshimura et al, Tetrahedron Lett. 1999, 40, 1937-1940].
In this context, a preparation route to compounds of the formula III is described
in which R3 and R4 represent alkyl, silyl or acyl, and R5 represents acyl (Scheme 1).

The starting material in Scheme 1 is the commercially available 1,2:5,6-di-O-isopropylidene-α-D-allofuranose (A1) which for its part can be obtained in four steps from D-glucose [D. C. Baker et al, Carbohydr. Res. 1972, 24, 192-197]. Thus, compound II can be obtained in a total of 14 chemical steps starting with 1,2:5,6-di-O-isopropylidene-α-D-allofuranose or 18 chemical steps starting with D-glucose. Here, compound II is obtained as an anomeric mixture consisting of IIα and IIβ. The literature does not make any statements about the ratio IIα/IIβ [Y. Yoshimura et al, J. Org. Chem. 1999, 64, 7912-7920]. Laboratory experiments carried out by the applicant for preparing II analogously to the literature procedure starting with A14 gave IIα/IIβ mixtures of 1:1 to 3:2.
A disadvantage of the process, known from the prior art, for preparing the compound II is the large number of chemical steps required, which makes the practice of the process on an industrial scale considerably more difficult. Furthermore, in particular when the synthesis is carried out on an industrial scale, there are the following difficulties and problems:                The process comprises at least five preparative chromatographic separations (prep-HPLC).        The intermediates A6, A7, A9, A12 are unstable.        Handling of the viscous liquids in stages A2, A3, A4, A8, A9, A11, A12 is difficult.        The compound A6 dissolves only very slowly in methanol. In the presence of sodium methoxide (NaOMe), a nucleophilic substitution of the mesylate group by a methoxy group in the compound A7 takes place as a side-reaction. Formation of by-product takes place in particular when the reaction is carried out on a relatively large scale.        After cleavage of the isopropylidene group, trifluoroacetic acid (TFA) has to be distilled off under reduced pressure since other alternatives for work-up result in a large formation of side-product. On an industrial scale, this is associated with considerable difficulties.        
Owing to the long synthesis sequence and the fact that some of its steps cannot be scaled up, or only with considerable expense, the process shown in Scheme 1 is not suitable for the industrial commercial preparation of the compound II.
An alternative for preparing the compound II is described in WO 2007/068113 and the literature associated therewith [J. K. Watts et al, J. Org. Chem. 2006, 71, 921-925] and is summarized here in Scheme 2.

Here, in the last synthesis step, a mixture of the anomeric acetates II in a ratio IIα/IIβ of from 1:2 to 1:14 is obtained [see also J. K. Watts et al, J. Org. Chem. 2006, 71, 921-925]. The by-product B8 is removed by column chromatography.
Compound B1 can be prepared in 6 steps from L-xylose (which does not occur in nature) [J. K. Watts et al, J. Org. Chem. 2006, 71, 921-925]. Thus, compound II can be prepared in a total of 13 chemical steps from L-lyxose.
A particular disadvantage of this synthesis alternative is due to the fact that the starting material L-lyxose is expensive and very little is commercially available for a synthesis on an industrial scale.
Furthermore, in particular when the synthesis is carried out on an industrial scale, there are the following difficulties and problems:                On each synthesis stage, complicated protective group transformations and in each case a chromatographic purification have to be carried out.        The use of liquid ammonia and elemental lithium at very low temperatures (step B2).        Introduction and removal of a particular silyl protective group which has a high molar mass and is difficult to obtain commercially (steps B3 and B5).        Use of DAST as fluorinating agent. In addition to the fact that DAST is difficult to obtain, safety concerns (handling temperature, decomposition of DAST in an exothermal reaction with formation of gas) play an important role in the scale-up of this reaction (step 4).        The use of ozone at very low temperatures (B7).        High temperatures (110° C.) during the Pummerer rearrangement and the formation of about 20% of by-product B8 [J. K. Watts et al, J. Org. Chem. 2006, 71, 921-925].        
Owing to the difficulties, described here, in the individual steps of the synthesis, which render scale-up difficult or impossible, and owing to the limited availability of the starting material, the process shown in Scheme 2 is likewise not very suitable for the industrial commercial preparation of the compound II.